Here at MAIL, We develop computational frameworks that enable existing complex/deep models to be more suitable for practical uses. We focus on improving the following aspects of existing models: (i) training/inference, (ii) realistic assumptions, (iii) algorithmic robustness, and (iv) efficiency in constrained settings. Most of our ML/AI solutions center around large-scale approaches that have low computational complexity and require less human effort.

Selected Press Coverage: khoahocphattrien , Thanh Nien, VnExpress, BaoDauTu, DanTri, Vietnam.vn, Vietnam.vn, Yahoo Finance, Benzinga, Macau Business, Taiwan News, TNGlobal, VinGroup…

Research Interests

Our research focuses on understanding the practical limits of using existing ML methods in the real-world. Essential, we seek answers to the following question: How to make ML models simpler & reliable to use in constrained settings? Simplicity refers to the ability to (i) build or implement the method easily, (ii) execute the deployed model efficiently, and (iii) evolve the deployed model with less effort. Reliability relates to (i) whether we can rely on the model to solve the intended task well, (ii) whether this performance is preserved under frequently perturbed environments in practice such as data corruptions or distributional changes, and (iii) whether the model is resilient to (i.e., its performance is not significantly affected by) various forms of security attacks such as adversarial examples and causal attacks. In this sense, we believe that many existing ML methods, including those with complex deep neural networks, are reliable but not yet easy-to-use because they do not satisfy various constraints seen in real-world applications. We also strongly believe the effort to answer this question will help us truly realize the potential of AI/ML methodology in practice.

Our goal, therefore, is to develop computational frameworks that enable existing complex/deep models to be more suitable for practical uses. We focus on improving the following aspects of existing models: (i) training/inference, (ii) realistic assumptions, (iii) algorithmic robustness, and (iv) efficiency in constrained settings. Most of our ML/AI solutions center around generative-based approaches that have low computational complexity and require less human effort. Currently, our research activities include, but not limited to, the following themes (with selected publications):

Trustworthy Machine Learning

We focus in several issues such as security vulnerabilities in training data, training process, and model inference, robust predictions under missing data and OOD scenarios, and vulnerabilities in Federated Learning systems.

• Overcoming Catastrophic Forgetting in Federated Class-Incremental Learning via Federated Global Twin Generator (2024 by Nguyen et al)
• Estimating Uncertainties of Multimodal Models with Missing Modalities (2024 Nguyen et al.)
• Non-Cooperative Backdoor Attacks in Federated Learning: A New Threat Landscape (2024 by Nguyen et al.)
• Everyone Can Attack: Repurpose Lossy Compression as a Natural Backdoor Attack (2024 by Yang et al.)
• Synthesizing Physical Backdoor Datasets: An Automated Framework Leveraging Deep Generative Models (2024 by Yang et al.)
• Flatness-aware Sequential Learning Generates Resilient Backdoors (ECCV 2024 by Pham et al.)
• Data Poisoning Quantization Backdoor Attack (ECCV 2024 by Huynh et al.)
• Composite Concept Extraction through Backdooring (ICPR* 2024 by Ghosh et al.)
• Fooling the Textual Fooler via Randomizing Latent Representations (ACL 2024 by Hoang et al.)
• Understanding the Robustness of Randomized Feature Defense Against Query-Based Adversarial Attacks (ICLR 2024 by Nguyen et al.)
• Backdoor attacks and defenses in federated learning: Survey, challenges and future research directions (EAAI 2024 by Nguyen et al.)
• Iba: Towards irreversible backdoor attacks in federated learning (NeurIPS 2023 by Nguyen et al.)
• A Cosine Similarity-based Method for Out-of-Distribution Detection (ICML-W 2023 by Nguyen et al.)
• Clean-label Backdoor Attacks by Selectively Poisoning with Limited Information from Target Class (NeurIPS-W 2023 by Nguyen et al.)
• Defending backdoor attacks on vision transformer via patch processing (AAAI 2023 by Doan et al.)
• Marksman Backdoor: Backdoor Attacks with Arbitrary Target Class (NeurIPS 2022 by Doan et al.)
• Backdoor Attack with Imperceptible Input and Latent Modification (NeurIPS 2021 by Doan et al.)
• LIRA: Learnable, Imperceptible and Robust Backdoor Attacks (ICCV 2021 by Doan et al.)

Information Retrieval and Applications

We focus in hashing-based methods and retrieval with complex, non-metric measures.

• Cooperative Learning of Multipurpose Descriptor and Contrastive Pair Generator via Variational MCMC Teaching for Supervised Image Hashing (2024 by Doan et al)
• Asymmetric Hashing for Fast Ranking via Neural Network Measures (SIGIR 2023 by Doan et al.)
• One Loss for Quantization: Deep Hashing with Discrete Wasserstein Distributional Matching (CVPR 2022 by Doan et al.)
• Generative Hashing Network (ACCV 2022 by Doan et al.)
• Interpretable Graph Similarity Computation via Differentiable Optimal Alignment of Node Embeddings (SIGIR 2021 by Doan et al.)
• Image Hashing by Minimizing Discrete Component-wise Wasserstein Distance (2021 by Doan et al.)
• Efficient Implicit Unsupervised Text Hashing using Adversarial Autoencoder (WWW 2020 by Doan et al.)

Efficient Generative AI

We focus on developing low-resource algorithms for generative AIs.

• Sparse Watermarking in LLMs with Enhanced Text Quality (2024 by Hoang et al.)
• Fair Generation in LLMs with RAG (Chu et al.)
• Reward Over-optimization in Direct Alignment Algorithms with Adaptive Learning (2024 by Nguyen et al.)
• Predictive Concept Attribution in Difussion Models (by Nguyen et al.)
• Synthesizing Physical Backdoor Datasets: An Automated Framework Leveraging Deep Generative Models (2024 by Yang et al.)
• Image Generation Via Minimizing Frechet Distance in Discriminator Feature Space (2021 by Doan et al.)

Low-resource Machine Learning

Currently, a large part of our research focuses on solving various societal challenges with AI in Vietnam. We’re working on problems such as low-resource NLP, low-resource remote-sensing predictive algorithms, cross-cultural language understanding, and visual question answering algorithms for medical domain. For more information, please visit the Center for Envrionmental Intelligence (CEI, where Prof. Khoa D. Doan is currently the Environment Monitoring Lab Director) and VinUni-Illinois Smart Health Center (VISHC, where Prof. Khoa D. Doan is currently the Associate Director)

VinUni-Illinois Smart Health Center (VISHC) – VISHC is open to collaborate with all researchers and research/industry institutions in Vietnam and around the world. VISHC aims to solve various healthcare related challenges with translational and innovative research. Led by Prof. Minh Do and Prof. Helen Nguyen at UIUC, and Prof. Khoa D Doan at VinUni (who leads MAIL-Research), the VISHC’s team comprises of world-renowned researchers and talented PhD/Master Students, Research Assistants and Postdocs. Please reach out via email for any collaboration opportunities.

Center for Environmental Intelligence (CEI) – MAIL-Research is a member of CEI. CEI represents a pioneering initiative at the intersection of advanced technology, environmental science, and interdisciplinary research and is open for collaboration. Led by Prof. Laurent El Ghaoui, CEI aims to address critical global sustainability challenges with innovative approaches based on AI. Please reach out via email for any collaboration opportunities, especially those related to environmental monitoring.